classification of building function using available ...€¦ · 2.3 foursquare launched in 2009,...

CLASSIFICATION OF BUILDING FUNCTION USING AVAILABLE SOURCES OF VGI

C. C. Fonte1,2∗, M. Minghini3, V. Antoniou4, J. Patriarca2, L. See5

1 Department of Mathematics, University of Coimbra, Coimbra, Portugal - [email protected] Institute for Systems Engineering and Computers at Coimbra (INESCC), Coimbra, Portugal - [email protected]

3 Department of Civil and Environmental Engineering, Politecnico di Milano, Milan, Italy - [email protected] Geographic Directorate, Hellenic Army General Staff, Athens, Greece - [email protected]

5 International Institute for Applied Systems Analysis, Laxenburg, Austria - [email protected]

Commission IV, WG IV/4

KEY WORDS: Volunteered Geographic Information, OpenStreetMap, Facebook, Foursquare, Building Function, Automated Classi-fication

ABSTRACT:

This paper examines the feasibility of using data from OpenStreetMap (OSM), Facebook and Foursquare as a source of informationon the function of buildings. Such information is rarely openly available and if available, would vary between cities by nomenclature,making comparisons between places difficult. Volunteered Geographic Information (VGI) including data from social media representsnew potential sources of building function data that have not yet been exploited for this purpose. Using a part of the city of Milan asthe study area, building data from OSM and points of interest (POIs) from OSM, Facebook and Foursquare were extracted to derive thebuilding function. This resulted in the classification of building function for more than 80% of the buildings and demonstrated that bothFacebook and Foursquare can complement the building function derived from OSM, helping to fill in missing gaps. This preliminarystudy has demonstrated the potential of this approach for deriving building function information from open data in a simple way yetstill requires independent validation with alternative sources as well as extension to other areas that have different amounts of OSM andsocial media coverage.

1. INTRODUCTION

Information on the function of buildings, i.e. whether a buildingis commercial, residential, industrial, mixed use, etc., is some-times recorded by local authorities or by national mapping agen-cies. However, such data are rarely openly available, yet they areincredibly valuable for a number of different applications rangingfrom the determination of energy demand (Caputo et al., 2013),evaluation of contributions to greenhouse gas emissions (Luconet al., 2014) and as inputs to urban climate and energy balancemodelling (Tornay et al., 2017). Furthermore, data on buildingfunction can be used in risk assessments of exposure, urban plan-ning applications, smart cities development and the efficient allo-cation of resources.

In lieu of actual data from authoritative databases, we can turnto Volunteered Geographic Information (VGI) (Goodchild, 2007)as a potential source of building function data. OpenStreetMap(OSM), which is one of the most popular and successful exam-ples of VGI (Mooney and Minghini, 2017), contains abundantinformation on buildings, including tags that can be used to de-rive the building function. OSM also contains Points of Interest(POIs), which can be used to complement the building informa-tion found in OSM. In addition to OSM, passive or ambient VGIfrom social media can be mined for information on building func-tion. For example, Facebook has many sites devoted to placesthat contain a geographic location. Foursquare is a mobile appli-cation that connects people with geographic locations, includingbuildings that have specific functions such as a restaurant or cafe.Together these different sources can be integrated to produce a

∗Corresponding author

building function data set, which has not been previously investi-gated. Moreover, such an approach could prove valuable in citieswhere databases on building function do not currently exist.

Hence this paper aims to demonstrate how data extracted fromdifferent sources of VGI can be combined to produce a data set onbuilding function. This approach is applied to an area within thecity of Milan. After brief descriptions of the main VGI sourcesused in this study, i.e. OSM, Facebook and Foursquare, the mainworkflow is outlined, from data extraction to integration into asingle building function database. The results are then presentedshowing the contribution of each VGI source in mapping build-ing function for the study area. Finally, limitations are outlinedfollowed by plans for further work in this area.

2. STUDY AREA AND VGI DATA SETS

The area selected for the study is a 2km x 2km section corre-sponding to the city center of Milan, Italy (see Figure 1). Thearea is characterized by a high density of buildings, includingcommercial activities such as shops and company offices, accom-modation, public buildings and government offices, churches,schools and universities, residential buildings and even a cas-tle (Sforzesco Castle). To automatically classify the function ofthese buildings, three VGI sources were exploited: OSM, Face-book and Foursquare. The following subsections describe eachof these sources in more detail.

2.1 OpenStreetMap

OpenStreetMap (OSM, https://www.openstreetmap.org) is themost popular VGI project to date. Started in 2004 as a crowd-sourcing project for road mapping, its focus was then extended

The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-4, 2018 ISPRS TC IV Mid-term Symposium “3D Spatial Information Science – The Engine of Change”, 1–5 October 2018, Delft, The Netherlands

This contribution has been peer-reviewed. https://doi.org/10.5194/isprs-archives-XLII-4-209-2018 | © Authors 2018. CC BY 4.0 License.

209

https://www.openstreetmap.org

Figure 1. Study area located in Milan, Italy. Map data:© OpenStreetMap contributors

to any physical object located on the Earth’s surface. As aresult, OSM is currently the largest, most diverse, most com-plete and most up-to-date geospatial database of the world. Dueto the richness of its content and the free and open accessgranted by its license, OSM is frequently studied by researchersand used heavily in cartographic, business, educational, gov-ernmental, humanitarian and leisure applications (Mooney andMinghini, 2017). The OSM database consists of vector fea-tures, i.e. geometric data types and related attributes. Ge-ometric data types consist of nodes, which are used to de-scribe objects that are points; ways, used to describe both lin-ear and polygon objects; and relations, which describe rela-tionships between two or more nodes, ways and/or other re-lations (https://wiki.openstreetmap.org/wiki/Elements). The at-tributes associated with the geometric data types are calledtags and consist of the combination of a key (defining the ob-ject property) and a value (specifying the object value for thatproperty). Each OSM object must have at least one tag, butthere is no limit to the number of tags an object can have(https://wiki.openstreetmap.org/wiki/Tags). The OSM mappingguidelines, which list the tags to be used for any specific ob-ject, have been agreed upon over the years by the OSM com-munity and are maintained on the Map Features wiki page(https://wiki.openstreetmap.org/wiki/Map Features). The evolu-tion of these guidelines over time has been recently studied froma data quality perspective (Antoniou and Skopeliti, 2017). Statis-tics on the current usage of each OSM tag are provided by thepopular Taginfo service (https://taginfo.openstreetmap.org).

2.2 Facebook

Facebook (http://facebook.com) is a digital platform that allowsits users to communicate with each other and share informationin multiple formats, including text, images and video, related toany topic of human activity. This massive social network hasmillions of users spread across all continents. Thus, it providesan enormous amount of information that can be used in mul-tiple applications. Among the diversified uses of Facebook, itcan be used to identify physical entities, such as buildings withsome type of use. Facebook also provides the explicit location

(with coordinates) of these entities as well as other useful infor-mation, such as functionality, services provided, profile of theusers or the quality of the services. Facebook provides a GraphAPI (https://developers.facebook.com/docs/graph-api), which of-fers tools for searching and downloading data. This digital socialnetwork is organized as a graph: ”nodes” are individual objects,such as a user, a photograph, a publication or a comment; and”arcs” are links associating a user with a publication, photographor place. This makes the search for Facebook data challengingbecause it forces the use of stratified search strategies. For ex-ample, if the goal is to identify the user ratings of restaurant ser-vices located in a particular geographical area, the first step isto identify and locate each service. Only after that is it possibleto establish, through a new survey, a link between those nodesand other elements with relevant information (e.g. publicationsor comments).

2.3 Foursquare

Launched in 2009, Foursquare (https://foursquare.com) is a mo-bile app offering location-based services to record consumer ex-periences and business solutions. Until 2014, it was primarily asocial networking tool through which users could share their cur-rent location (e.g. a restaurant, a museum or a shop) with friendsthrough the so-called “check-in” function. In 2014, the focus ofthe Foursquare app changed to become primarily a local searchand discovery tool (https://en.wikipedia.org/wiki/Foursquare).Using a proprietary technology to detect the location of users,Foursquare allows them to search for places of interest in theirsurroundings and displays personalized recommendations basedon the time of day, user check-in history, their tastes and previ-ous ratings. Users can write tips, i.e. short messages about alocation, which can also include a photograph. Tips can be likedand saved by any user. A user writing quality tips, i.e. with sev-eral likes and saves, is awarded with experience points, whichmakes their content more prominent. Foursquare allows usersto specify their tastes, in particular food items, styles of cuisineor environmental aspects; these are matched to the available tipsat surrounding venues to provide customized searches and rec-ommend places that match the user tastes. Finally, Foursquareallows users to rate places of interest by answering questions. Inaddition to determining the popularity of a venue, ratings are alsoused to get complete information about a venue (e.g. whethera restaurant takes credit cards). Foursquare, which currently in-cludes 105 million venues around the world, is one of the moststudied examples of location-based services that have adopted agamification approach (Frith, 2013).

3. METHODOLOGY

The target classes selected for building functions are the follow-ing: residential, commercial, industrial, educational, healthcare,cultural, entertainment, religious, accommodation, safety author-ities, agricultural, civic amenities/government offices, transporta-tion, non-governmental/non-profit, and military. The workflowadopted to associate one of these classes to the buildings in Mi-lan first requires the extraction of the building data (i.e. poly-gons) from OSM, while the POIs (i.e. points) were extractedfrom OSM, Facebook and Foursquare. The OSM buildings andthe OSM, Facebook and Foursquare POIs were then separatelymapped onto the building function classification. Figure 2 showsthe workflow used to associate each POI to the correspondingbuilding, which requires checking whether the POI was inside abuilding or not; if it was outside but within a maximum distance



210

https://wiki.openstreetmap.org/wiki/Elements

https://wiki.openstreetmap.org/wiki/Tags

https://wiki.openstreetmap.org/wiki/Map_Features

https://taginfo.openstreetmap.org

(5m) from a building, it was assigned to the closest building. Fi-nally, the classification results are computed for each single VGIsource and for their combination. Each of these steps is discussedin more detail in the following subsections.

Figure 2. Workflow used to assigned the POI function to thebuildings

3.1 Extraction and classification of OSM buildings

As mentioned above, both buildings and POIs were extractedfrom OSM. The extraction was performed in July 2018 us-ing the HOT Export Tool (https://export.hotosm.org/en/v3),which relies on a version of the OSM database updated ev-ery few minutes and allows users to perform customizeddownloads in terms of the object tags. The main OSMkey defining the building function, which is attributed to thepolygon outlining the building (a ”way” in the OSM datamodel), is building, which can assume all the values listed inhttps://wiki.openstreetmap.org/wiki/Key:building. Some valuesof the building key do not specify any building function; the mostfrequently used key is yes, which simply defines the presence of abuilding and is adopted when the building is digitized from satel-lite imagery; this usually makes it impossible to define its func-tion. Table 1 shows the values of the building key associated withthe selected classes of building function.

In some cases, the tag building=yes is combined with another tagthat can reveal its function, e.g. amenity=restaurant to indicatethat the whole building is used as a restaurant. These additionaltags, not included in Table 1 and mainly corresponding to thekeys amenity, shop, tourism, office, historic and man made, werealso considered for deriving the building function.

3.2 Extraction and classification of OSM POIs

The OSM POIs (consisting of nodes in the OSM data model),were also extracted in July 2018 using the HOT Export Tool.The OSM keys used to associate the POIs with the selected tar-get classes are amenity, shop, tourism, office, industrial, power,leisure, historic, man made, healthcare, social facility and mil-itary. For brevity, the full table associating the OSM POI tags

Building function Values of the building key

residential

apartments, house, detached,residential, dormitory, terrace,houseboat, bungalow, static caravan,carport, conservatory, garage, garages,garbage shed, shed

commercial commercial, office, retail, kiosk, cabin

industrialindustrial, warehouse, bakehouse,digester, hangar, transformer tower,service, water tower, pumping station

educational kindergarten, school, universityhealthcare hospitalcultural pavilion, castleentertainment stadium, grandstand, riding hall

religiousreligious, cathedral, chapel, church,mosque, temple, synagogue, shrine,monastery

accommodation hotel, hutsafety authorities -

agriculturalfarm, farm auxiliary, barn, cowshed,greenhouse, stable, sty

civic amenities/government offices

public, civic

transportationtrain station, transportation,aerodome, bridge, parking,ferry terminal

non-governmental/non-profit

-

military bunker

no functionyes, construction, roof, ruins,collapsed

Table 1. Values of the OSM building key associated with eachselected class of building function

to the building functions is not included here. By way of an ex-ample, a POI is classified as belonging to a religious building if ithas at least one of the following tags: amenity=place of worship,office=religion, man made=campanile, historic=church and his-toric=wayside shrine. It should be noted that according to theMap Features wiki page, many values for the keys considered forPOIs do not refer to building features and were thus excludedfrom the analysis. Examples are the tag historic=memorial, usedto identify memorials, and the tag amenity=taxi, used to mark aplace where taxis wait for passengers.

3.3 Extraction and classification of Facebook POIs

The extraction of POIs from Facebook was accomplished usingthe “Place Search” endpoint of the Graph Facebook API and theGeoPandas Python Package. Given one area of interest (definedby a central point and a maximum distance from it in meters), the“Place Search” was used to retrieve the location of the Facebookpages of type Place and some of their attributes given by the API,including name, category list, about and description. To use thethis tool, the search area needs to be a circle. As the study areais not circular, a buffer was defined around a point located inthe center of the region under analysis, ensuring that the buffercontains the entire study area. Only one request was requiredto extract all available data within the search area limits. Thedata available in the extracted attributes were used to assign oneof the target classes selected to associate building functions witheach POI, using keywords. For example, store, shop or restaurantwere assigned the class commercial while ballroom or club would



211

https://export.hotosm.org/en/v3

https://wiki.openstreetmap.org/wiki/Key:building

be assigned to the class entertainment.

3.4 Extraction and classification of Foursquare POIs

The Foursquare POIs were also extracted in July 2018 usingthe Personal/Non-Commercial Places API provided for free byFoursquare (https://developer.foursquare.com/places-api). The“Search for Venues” request was used in order to retrieve thepoints. Due to the API limitations, a maximum of 50 POIs (i.e.Foursquare venues) could be returned in each request. Thus,given the bounding box of the study area, several requests weresubmitted for different locations inside the study area. The re-quests were densely allocated so that no venues was omitted. Thisprocess resulted in multiple entries for some venues, which had tobe removed before the analysis. Although the API provides richinformation for each venue, the focus here was limited to their ge-ometry, name and classification. The latter can be derived fromthe content of the category field. Again, some values occurringin this field are not related to buildings and were thus excludedfrom the analysis, e.g. park, road and scenic lookout. A mappingbetween the values found in these fields was then made to the tar-get classes for building functions selected in this study. The fullmapping is not included here due to space limitations, but someexamples are the classification of the categories restaurant, baror shop as commercial and the classification of Art Museum tocultural or church to religious.

3.5 Association of POIs with buildings

To assign the POI characteristics to the building function classes,it is necessary to establish a relation between the POIs and thebuilding polygons. This was done using the spatial location ofthe POIs. If a POI was located inside the polygon of a build-ing, it was considered that its characteristics could be associatedwith the building. However, in many situations, volunteers do notplace a POI inside the building polygon, but close to the buildingentrance instead; this highlights access to what the POI refers to,such as a particular shop. Therefore, a proximity analysis wasadditionally used to associate the POIs to the buildings. For thepurposes of this study, if a POI is within a 5m buffer of a par-ticular building (and not located inside any other building), theinformation from that POI was then associated with that build-ing. Tests were made considering different distances. The 5mvalue was used in the analysis because it proved to be a goodcompromise between considering POIs that are, in fact, not as-sociated with the building and discarding useful POIs that arefurther away from buildings.

This approach for associating POIs with the buildings will, inmany cases, associate several POIs to one building. As thesePOIs may have different functions, an additional building func-tion category was added to the analysis called multiple, which isassigned to buildings that are either associated with several POIswith different functions, or the POI functions are different fromthe function that is extracted from the building tag.

4. RESULTS

4.1 Extraction and classification of OSM buildings

In total, 2902 buildings were extracted for the area depicted inFigure 1. In a previous study, a comparison against the authori-tative building data set showed that the OSM buildings located inthe city center of Milan are characterized by a very high com-pleteness and positional accuracy (Brovelli et al., 2016). The

classification of these buildings into the selected target functionswas performed according to the matching rules of Table 1 and isdisplayed in Figure 3. Using only the tags associated with theOSM building polygons, 1249 buildings (43% of the total) couldbe associated with a specific class of building function. The re-maining 1653 buildings (57% of the total), represented in gray inFigure 3, remained unclassified. The fraction of classified build-ings associated with the target functions are the following: res-idential (75%), commercial (15%), religious (4%), educational(2%), healthcare (1%), civic amenities/government offices (1%).Accommodation, cultural, industrial and transportation appearwith percentages lower than 1%, while no buildings with enter-tainment, safety authorities, agricultural, non-governmental/non-profit and military functions were found.

Figure 3. Classification of building function using only theattributes from the OSM building polygons. Map data:

© OpenStreetMap contributors

4.2 Extraction and classification of POIs

Table 2 shows the total number of POIs extracted from OSM,Facebook and Foursquare for the study area. For each of the threesources, the table also shows the number and fraction of POIslocated inside the OSM buildings, and the number and percentageof POIs located within a distance of 5m from at least one building.

Data sourceNumber ofextractedPOIs

Number ofPOIs insidebuildings

Number ofPOIswithin 5mof buildings

OSM 2197 1978 (90%) 128 (6%)Facebook 58 30 (52%) 7 (12%)Foursquare 312 91 (29%) 75 (24%)

Table 2. Total number of POIs extracted from OSM, Foursquareand Facebook, and the number and fraction of POIs located

inside and within a distance of 5m from buildings



212

4.2.1 OSM POIs: 2197 POIs were extracted from OSM us-ing the methodology outlined in Section 3.2. The tags of 2193POIs could be associated with a building function, while only 4remained unclassified, corresponding to the tag historic=ruins,which describes the remains of buildings fallen into partial orcomplete disrepair. The percentage of classified POIs associatedwith the target functions are the following: commercial (86%),civic amenities/government offices (6%), accommodation (3%),cultural (3%) and educational (1%). The healthcare, entertain-ment, religious, safety authorities and non-governmental/non-profit functions appear with percentages lower than 1%, while noPOIs were found with residential, industrial, agricultural, trans-portation and military functions. From these, 90% were locatedinside buildings (see Table 2) and 6% within 5m of buildings.The remaining 4% were further away from buildings and were,therefore, not considered in this analysis.

Several of the extracted POIs, some of them with different classi-fications in terms of function, are located inside the same build-ing. Figure 4 shows the number of POIs extracted from OSMlocated inside each building, while Figure 5 shows the number ofdifferent functions that can be associated with the building usingthe OSM POIs that are located inside the building. Even thoughsome buildings have many POIs (with a maximum of 30 POIsinside one building), the number of different functions obtainedfrom this source is in most cases only one, and only a few build-ings have three different function categories.

Figure 4. Number of POIs from OSM located inside eachbuilding. Map data: © OpenStreetMap contributors

4.2.2 Facebook POIs: Only 58 POIs were extracted fromFacebook (see Table 2). From these, only three could not beassociated with a building function. One corresponds to thecity center of Milan, another to a ”Plaza” and the other toa ”historical place”. The other POIs that are associated withthe target building functions considered here are split by com-mercial (50%), cultural (14%), entertainment (13%), religious(5%), non-governmental/non-profit (5%), educational (4%), civicamenities/government offices (4%) and the remaining functions

Figure 5. Number of different functions associated withbuildings using the OSM POIs located inside each building.

Map data: © OpenStreetMap contributors

2%. From the extracted POIs, only 64% were associated withbuildings (52% inside and 12% in the 5m vicinity) (Table 2).

4.2.3 Foursquare POIs: From Foursquare, 312 POIs wereextracted (see Table 2). From these, 11% could not be associ-ated with any building functions as they provided data about thelocation of plazas, gardens or parks. For the POIs with associatedbuilding functions, the great majority have the function commer-cial (85%), followed by accommodation (8%), cultural (5%) andentertainment (2%). All other functions were not found in theFoursquare POIs, except for religious, with only one POI (0.4%).Only 53% of the POIs extracted from Foursquare were insidebuildings (29%) or in their 5m vicinity (24%), and therefore onlythese were used in this study.

4.3 Association of POIs with buildings

Figure 6 shows the building function classification obtained withthe POIs extracted from OSM (inside and in the 5m proximity).It can be seen that many buildings (1019, corresponding to 35%of the buildings in the study area) have at least one function as-signed.

Figure 7 shows the building classification obtained with the POIsextracted from Facebook. It can be seen that only 32 buildingswere classified (1% of the buildings in the study area). However,a few buildings that did not have an assigned function from OSMhave now been assigned to one or multiple functions.

Figure 8 shows the building classification obtained with the POIsextracted from Foursquare. These POIs allowed functions to beassigned to 163 buildings (6% of the buildings in the study area).It can be seen that multiple functions are associated with some ofthe buildings.

Figure 9 shows all of the buildings in the study area that were as-signed a function using either the OSM building tags or the POIs



213

Figure 6. Classification of building function using the POIsderived only from OSM. Map data: © OpenStreetMap

contributors

Figure 7. Classification of building function using the POIsderived only from Facebook. Map data: © OpenStreetMap

contributors

extracted from the 3 data sources. This resulted in 2382 clas-sified buildings (82% of all buildings in the study area). Fromthese, 867 were classified as having multiple functions. For ex-ample, the Sforzesco Castle located in the northwestern part ofthe study area was classified as having multiple functions: cul-

tural and civic amenities/governmental offices. The cathedral lo-cated close to the center of the study area (also classified withmultiple functions) had the following two functions assigned bythe POIs: religious and cultural.

Figure 8. Classification of building function using the POIsderived only from Foursquare. Map data: © OpenStreetMap

contributors

Figure 9. Classification of building function using the OSMbuildings and the POIs derived from OSM, Facebook and

Foursquare. Map data: © OpenStreetMap contributors



214

5. CONCLUSIONS

This study has shown how different sources of VGI includingdata derived from social media can be used to classify buildingsaccording to their function. More than 80% of the buildings inthe study area could be classified using a detailed set of build-ing function classes. This information is rarely openly availableyet has many potential applications.Such an approach provides asimple yet powerful way of extracting this information.

Most of the data about building function was extracted fromOSM, either from the building tags or the POIs. However, theaddition of data obtained from Facebook and Foursquare allowedfunctions to be assigned to some buildings that were not classi-fied with OSM data, filling in some gaps and demonstrating thecomplementarity of these different data sources. In contrast tothe OSM building tags, most POIs identified buildings with somekind of commercial activity. Moreover, POIs do not provide dataon residential buildings so to obtain this type of function, OSMbuilding tags are the only source of information. We also ac-knowledge that the building function characterization considersactivities that can be associated with parts of a building, e.g. theground floor, and does not currently allow the specification ofdifferent building functions per building section, e.g. commer-cial shops in the ground floor with residential locations above, asno information about the building height or the separation of usewithin the building is considered.

At present this approach has not been validated using an in-dependent data source. In the future, an accuracy assessmentwill be undertaken using data from authoritative databases orother independent sources. The latter could include photographsfrom online repositories such as Panoramio and Flickr, GoogleStreetView, Mapillary or field visits. In addition to validation, wecould also use these latter sources of information to complementor enrich the analysis, e.g. they could provide missing data onbuilding function, thereby filling in missing gaps.

In the future we will also map the buildings with multiple func-tions to each of the functions actually assigned to it as well asassigning a degree of confidence to each building and function.This value would be based on the number of data sources thatproduced the classification relative to all data sources used. Themore data sources that provide the same information, the higherthe confidence. This could also help to determine if the buildingreally has multiple functions or if a single function dominates.

Finally, we plan to apply this procedure to other cities in order toassess the usefulness of each source of data in other countries andregions that have different characteristics, e.g. different coverageand tagging behavior in OSM or greater use of social media indocumenting places.

ACKNOWLEDGEMENTS

The study has been partly supported by the the Portuguese Foun-dation for Science and Technology (FCT) under project grantUID/MULTI/ 00308/2013, the FP7 ERC project CrowdLand (No.617754), the Horizon2020 LandSense project (No. 689812) andthe URBAN-GEO BIG DATA, a Project of National Interest(PRIN) funded by the Italian Ministry of Education, Universityand Research (No. 20159CNLW8).

REFERENCES

Antoniou, V. and Skopeliti, A., 2017. The Impact of the Contribu-tion Micro-environment on Data Quality: The Case of OSM. In:G. Foody, L. See, S. Fritz, P. Mooney, A.-M. Olteanu-Raimond,C. C. Fonte and V. Antoniou (eds), Mapping and the Citizen Sen-sor, Ubiquity Press, London, pp. 165–196.

Brovelli, M. A., Minghini, M., Molinari, M. E. and Zamboni,G., 2016. In: The International Archives of the Photogrammetry,Remote Sensing and Spatial Information Sciences, Prague, CzechRepublic, Vol. XLI-B2, pp. 615–620.

Caputo, P., Gaia, C. and Zanotto, V., 2013. A methodology fordefining electricity demand in energy simulations referred to theItalian context. Energies 6(12), pp. 6274–6292.

Frith, J., 2013. Turning life into a game: Foursquare, gamifi-cation, and personal mobility. Mobile Media & Communication1(2), pp. 248–262.

Goodchild, M. F., 2007. Citizens as sensors: the world of volun-teered geography. GeoJournal 69(4), pp. 211–221.

Lucon, O., rge Vorsatz, D., Ahmed, A. Z., Akbari, H. and et al.,2014. Buildings. In: Climate Change 2014: Mitigation of Cli-mate Change. Contribution of Working Group III to the FifthAssessment Report of the Intergovernmental Panel on ClimateChange, Cambridge University Press, Cambridge, United King-dom and New York, NY, USA.

Mooney, P. and Minghini, M., 2017. A Review of Open-StreetMap Data. In: G. Foody, L. See, S. Fritz, P. Mooney, A.-M.Olteanu-Raimond, C. C. Fonte and V. Antoniou (eds), Mappingand the Citizen Sensor, Ubiquity Press, London, pp. 37–59.

Tornay, N., Schoetter, R., Bonhomme, M., Faraut, S. and Mas-son, V., 2017. GENIUS: A methodology to define a detaileddescription of buildings for urban climate and building energyconsumption simulations. Urban Climate 20, pp. 75–93.



215

classification of building function using available ...€¦ · 2.3 foursquare launched in 2009,...

Documents